New Formula Gauges Disease-Free Colorectal Survival
Elsevier Global Medical News. 2011 Apr 1, P Wendling
DETROIT (EGMN) - A simple formula that incorporates tumor area and percentage of positive lymph nodes helps predict 5-year disease-free survival in colorectal cancer, and may be used to guide treatment, a new study suggests.
"The formula is simple and inexpensive to apply," lead author Dr. Lisa Poritz said at the annual meeting of the Central Surgical Association. All the data are in the pathology report. "It does not require genetic or biologic tests that are expensive and not available throughout the country."
Although tumor size is not part of the widely used TNM (tumor, node, metastasis) staging system, Dr. Poritz and her colleagues hypothesized that it would be important in determining disease-free survival. To test that hypothesis, they used data from pathology reports to calculate the tumor area-to-node ratio (T:N) for 63 patients with stage III colorectal cancer (CRC) who underwent resection at the Hershey (Pa.) Medical Center from January 2000 to June 2008.
Specifically, tumor area is based on the two largest tumor measures multiplied together. The percentage of positive nodes is based on the number of nodes that are positive divided by the total number of nodes harvested, then multiplied by 100. The T:N ratio is based on the tumor area divided by the percentage of positive nodes.
In all, 35 patients remained disease free at 5 years and 28 developed metastatic disease at a mean of 16.4 months, said Dr. Poritz, a colorectal surgeon at Penn State Hershey.
The mean T:N ratio was significantly higher (2.76) in the disease-free group, compared with 0.50 in the metastatic group (P = .0009), meaning that a higher ratio is a better prognostic sign, she said.
In a receiver operating characteristic analysis, T:N ratio demonstrated the highest correlation with 5-year disease-free survival at an area under the curve of 0.86 vs. 0.79 for percentage of positive nodes, 0.75 for number of positive nodes, and just 0.61 for tumor volume.
When the data were applied to a logistic regression curve, the beta0 was 1.682 and beta1 was 2.274. From these values, the risk of distant metastasis can be calculated for a specific T:N ratio. For example, a patient with a T:N ratio of 1.22 would have a 25% risk of developing metastatic disease in 5 years, compared with a 50% risk for a patient with a T:N ratio of 0.77 and a 75% risk for a patient with a T:N ratio of 0.26, Dr. Poritz said.
Notably, even after controlling for the critical factor of nodal stage, researchers found that the T:N ratio remained significant in logistic regression analysis at a P value of .008.
Dr. Poritz also presented two hypothetical patients to illustrate how the formula could be used to influence clinical decision making. Patient A had a small tumor with an area of 1.52 and 2 of 17 positive nodes. This resulted in a T:N ratio of 0.25 and a 75% risk of developing metastatic disease in 5 years. Patient B had a large tumor with an area of 28.3 and also 2 of 17 positive nodes. The T:N ratio would be 2.4, and the patient would have only a 2% chance of developing metastatic disease within 5 years.
"If you only looked at the number of nodes or [percentage of positive nodes], you might consider treating these patients in exactly the same manner, or at least consider their risk of developing metastatic cancer to be exactly the same," she said. "However, when you use our calculations, you can see that tumor volume is very important, and it changes the risk of developing metastatic disease dramatically in these patients. In this context, you might want to treat patient A more aggressively than patient B."
Invited discussant Dr. Anthony Senagore of the University of Southern California in Los Angeles asked whether data were available on the relationship between T stage and the size of the tumor, because "you seem to imply that what we historically believe about depth of penetration seems to be less important, and that it's the mass of the tumor that matters."
Dr. Poritz replied that depth of penetration - which is included in the TNM staging system - remains important, especially in stage II and node-negative disease when clinicians are trying to determine who should receive chemotherapy. However, the data would suggest that at least within stage III disease, tumor area may be more important than T stage, which has not been shown to correlate well with prognosis either in their study or others.
An audience member remarked that the researchers may be making too big a leap in suggesting that the formula could be used to guide therapy, and asked whether they have considered taking the findings to SWOG (formerly the Southwest Oncology Group) for further validation. Dr. Poritz said that they are in the process of applying the formula to a larger database, similar to that of SWOG.
Colon cancer was present in 21 patients in the disease-free group and in 19 patients in the metastatic group, with 14 cases of rectal cancer in each group. There were no differences between the two groups with regard to sex, age, tumor location, vascular or lymphatic invasion, or T stage. Perineural invasion was significantly more common in the metastatic group than in the disease-free group (11 cases vs. 2 cases), but Dr. Poritz urged caution in interpreting this because data on perineural invasion were not mentioned in 30 cases.
As expected, significantly more patients in the metastatic group than in the disease-free group had an N stage of 2 (18 cases vs. 6 cases) or N stage of 3 (1 case vs. 0). Similarly, the mean number of positive nodes was higher in the metastatic group than in the disease-free group at 6.1 vs. 2.3 (P = .0049), as was the percentage of positive nodes at 39.6% vs. 13% (P = .0006).
The study excluded patients who had inflammatory bowel disease, associated CRC, recurrent cancer, or perforated cancer, or who had received any neoadjuvant therapy or had died of unknown causes or causes other than cancer.
The authors reported no study support or conflicts of interest.
Comments